Search
Browse By Day
Browse By Time
Browse By Person
Browse By Policy Area
Browse By Session Type
Browse By Keyword
Browse Artificial Intelligence Presentations
Program Calendar
Sign In
Search Tips
Within the United States, both the federal government and the states have been introducing legislation related to the impacts of artificial intelligence (AI) technology for the past several years, with over 1,000 AI-related bills being introduced since January 2023. The policy goals of these bills vary greatly, including requirements on AI-related education, redress for AI-related harms, and efforts to stimulate innovation. Different states have taken different approaches, and it is often difficult to navigate the landscape of emerging legislation. This paper presents a framework and methodology for characterizing the content of AI-related legislation, as well as analysis of an initial dataset collected in early 2026. We first describe the development process for a questionnaire to be used by human analysts to characterize a piece of AI legislation. The questions in this survey span six categories: accountability and transparency, bias and discrimination, data protection, education, synthetic content, and labor. We then describe the coordination process for a cohort of over 20 undergraduate and graduate student policy analysts to apply this questionnaire to a dataset of over 100 AI-related bills across all fifty states, including criteria for prioritization of new bills to be evaluated. Finally, we analyze state-by-state variations in approaches to AI-related bills in this dataset. With this dataset, we are able to characterize variations in several key policy choices: the definitions of AI adopted, the introduction of civil remedies or criminal penalties, the imposition of requirements on AI developers, standards for AI-related education, data privacy practices, and more. We identify different points of convergence and divergence among the states in both the types of bills that are introduced and which bills are ultimately passed. In addition to the analysis in this paper, we will also publicly release the dataset so that other policy researchers can have a useful tool to explore their own questions regarding variations in AI legislation in the U.S.